Tuesday, October 8, 2013

Electoral Effects of the Shutdown

How will the shutdown change 2014 election results?

There are basically two ways a national swing can happen in Congressional races.

One is direct: voters shift from one party to the other. That one almost always happens against an incumbent party, generally against the party of the president. It's of course technically possible for it to go the other way, but it's highly unlikely. And the shutdown is especially unlikely to produce any such direct electoral action -- because even if voters really do care, and even if swing voters blame Republicans, it's unlikely that they're going to still be thinking about it by next November. So I wouldn't pay much attention at all to current polling for its predictive value (and I think I've already mentioned that I wouldn't pay any attention at all to the recent PPP poll in competitive Republican-held House districts; I think Sam Wang is totally wrong to believe that there's anything predictive there at all).

So a direct effect of voters punishing Republicans is highly unlikely (it's actually probably more likely that voters would punish Democrats if there are long-lasting economic effects, although it's probably true that elite opinion placing the blame squarely on the GOP probably would neutralize that).

However, there's also a potential indirect effect. Conventional wisdom in House races (and to a lesser extent Senate elections) can be self-fulfilling. If party actors on both sides believe that it's going to be a good cycle for one party, then that party will recruit better candidates, suffer fewer retirements in tough-to-defend districts, and have plenty of resources available. And since candidate quality is extremely important in House elections, that's going to swing seats in the favored party's direction -- even if voters don't have any connection to this "trend" at all.

There's a real chance that this indirect effect will really show up this time around, and now a bit of evidence -- reported in an excellent item from Greg Sargent -- that there are already some real candidate recruitment effects on the Democratic side. To be sure, DCCC claims should be taken with a grain of salt, especially with no actual candidate names attached. But then again the polls showing Democrats "winning" the shutdown are really only a couple of days old.

I'll note one other thing. The almost-universal interpretation of the 1995-1996 shutdowns also should work for Democrats, in two ways. First, it's one of the reasons that neutral reporters are likely to interpret this one the same way (just as reporters were likely to believe that a Republican landslide was possible or even likely during the 2010 cycle). And, second, party actors are, even beyond the reporting, probably set to interpret it that way.

On the other hand: it's also a midterm election, and "everyone" knows that midterm elections are bad for the party in the White House -- which means that any effects from the shutdown will be competing against effects from that other piece of conventional wisdom. Not only that, but there's probably a more general conventional wisdom right now that all midterms are good for Republicans, and all presidential years good for Democrats (it's mostly bunk, but what matter here is what people believe). So that, too, is something pushing all those bright young district attorneys, state senators, and local celebrities towards running this time if they are Republicans, and against it if they are Democrats.

All of which is to say that it's likely that the best Democrats can realistically hope for is for indirect effects of the shutdown to cancel out the other indirect effects out there and produce a push election, or maybe small gains. Nothing is impossible, of course, and if this winds up being significantly worse than the 1995-1996 shutdowns then we will be in uncharted territory of sorts. But I'd be very surprised by a large Democratic win in 2014.


  1. >Not only that, but there's probably a more general conventional wisdom right now that all midterms are good for Republicans, and all presidential years good for Democrats (it's mostly bunk, but what matter here is what people believe).

    Could you explain why you disagree with that conventional wisdom?

    1. For midterms being bad for Democrats--2006?

      For presidential years being good for Democrats--2004?

      Of course you can say that those are just two elections, and things have changed since then. But three elections--2008, 2010, and 2012--are hardly enough to justify the conventional wisdom that presidential election years belong to Democrats and midterms to Republicans. If the recession had developed a year later than it did, the Republicans could easily have won in 2008 and taken a terrible beating in 2010.

    2. The theory is a variant of an older theory in PS: "surge and decline." That theory argued that the winning president won because he got marginal supporters to the polls, and they rolloff more than others by the midterm. The theory still has adherents, but I think they are relatively few in number.

      The other source of the CW is the idea that Republicans are more consistent voters than Democrats are. There is some truth to this, but I don't think it's generally been a big enough effect to make a huge difference.

      Indies hold the balance in most elections, midterm or general. There are just fewer of them in midterms.

  2. By this rationale, the GOP House caucus could hold a ritual child-roasting in the Capitol rotunda in 2013 and there would be virtually no chance of a Democratic takeover in 2014 because voters will have forgotten all about it by then.

    1. I've learned not to laugh about things in The Onion because they eventually come true.


    2. Now I have to clean off my computer screen since I was drinking coffee when I got to "Obamastroid..."

  3. Here's my question, how do you explain the 2002 midterms then? Would you count 9/11 etc as being an event that was so big it was able to dominate the political landscape and produce a atypical incumbent victory? And if so, could a default or a really long shutdown qualify for a similar effect?

    1. I'd have to go back and refresh my memory, but off the top of my head: during the months in which candidates were making choices, the president had a >70% approval rating. That's a big deal!

      A massive recession which "everyone" blamed Republicans for might qualify. But people really do tend to blame the president (and his party) no matter what, so I wouldn't want to bet on it.

    2. Yet another argument for the basic superiority of European parliamentary democracy over a Madisonian system.

  4. Sam Wang has a great analysis for you: http://www.washingtonpost.com/blogs/wonkblog/wp/2013/10/08/republicans-could-lose-their-house-majority-because-of-the-shutdown/

    Also, see this excellent analysis by Frum: http://www.thedailybeast.com/articles/2013/10/08/seven-habits-of-highly-ineffective-political-parties-part-one.html

    1. Wang writes:

      "Usually, the president's party loses House seats at the midterm election, so the normal expectation would be for Republicans to gain seats. But exceptions can happen. In 1998, Democrats picked up five seats. That rather surprising result came after a rough few years, during which Speaker Newt Gingrich led the way to the last substantial shutdown, lasting 21 days."

      Wang here is trying to connect the 1995-6 shutdown with the 1998 midterms, but there's no reason to do so. It was the '96 election that immediately followed the shutdown, and the Republicans lost just 3 seats in the House--hardly a blowout, particularly in a presidential election year where they also lost the presidential election. Perhaps things will be different this time, but the point is that the '95 debacle doesn't provide any evidence that shutdowns have much if any effect on House races.

      It's generally acknowledged that it was the impeachment that led to the Dems' unprecedented (albeit small) gains in 1998. Blaming it on the shutdown that happened an election cycle earlier, and didn't seem to have much of an effect on that election, is simply confused. Weirdly, Ed Kilgore made this very same error recently when trying to refute Ramesh Ponnuru's claim that the '90s shutdown debacle didn't have much of an effect since it only led to a loss of a few seats: "The rationalization that the GOP 'lost only a few seats' reflects some serious amnesia. This was the only time in U.S. history that the party holding the White House for two consecutive terms gained House seats in the second midterm election."

      (As a side note, Kilgore isn't even correct that '98 was the only time the party controlling the White House gained seats in the second midterm. It happened in 1822.)

      One way or another, some commentators seem be to trying awfully hard to shoehorn the evidence to fit their own theories.

    2. There were a large number of retirements from the House in 1995-6--disproportionately from Democrats who found the body not to their liking any more now that Republicans were in control for the first time in decades (and their own seats no longer safe). Also, of course, a few left to run for higher office. Under those circumstances, it was hard for the Demcorats to make large gains. They did virtually tie the GOP in popular vote for the House, though, a substantial improvement from 1994.

      Some Democratic retirees who were replaced by Republicans in 1996: Tom Bevill (AL-07), Pete Geren (TX-12), Bill Brewster (OK-03), Sonny Montgomery (MS-03), Glen Browder (AL-03), Richard Durbin (Il-20), Tim Johnson (SD): also Cleo Fields of Louisiana was redistricted out of his seat.

      Had they run, most of these Democrats, having survived 1994, would probably have been favorites to win in 1996, and the Democratic gains would have looked somewhat more impressive. (Of course without Johnson and Durbin running for the Senate, the Democrats might have done worse there.)

      As for 1822, IMO opinion it doesn't count because the Democratic-Republicans had a virtual monopoly (all the presidential candidates in 1824 would consider themselves members of that party). The Federalists were all but dead.

    3. And Monroe had run unopposed in 1820.

    4. Yes I'm aware of all that, but it doesn't make Kilgore's claim correct (he said all of U.S. history). More importantly, his (and Wang's) larger confusion between the '96 and '98 cycles undermined his entire point. I also should note that the presumed distinction between first-term and second-term midterms may not be relevant. For the White House's party to gain seats in either situation is basically a case of Rare and Rarer. And since both situations have occurred in the last 15 years, maybe there is something different about recent times that makes it likelier than it was in the past.

    5. I have nothing to contribute to the discussion of the 1822 election cycle, but I am enjoying it.

  5. Economists are predicting a recession deeper than 2008 if US debts aren't paid on time. If so, the country will likely still be economically wounded, if not in deep recession, in 2014. Dem voters who normally would not bother with midterm elections will vote, as will sensible Rs, assuming there are any left. As 2012 showed, nationally the Rs are overmatched by the D constituency, which is growing. What that might mean in aggregate gains in the House I leave to political scientists to figure out.

    That's a theory of reasonable behavior. But we're entering dangerous territory, and all the academic theories may not apply. For me I do not see this as business as usual, governed by the usual factors. On top of the sequester and given other trends, default will likely be devastating for all but an even smaller group of, rich people. If it leads to inflation, the effects will be very visible. If it leads only to unemployment and poor people quietly suffering and dying at home, I expect that will take long enough that the media--owned by some of those rich people--will get bored and frightened.

    At best, this deliberate destruction by a handful of delusional congressmen aided and abetted by cowardly leadership could lead to a real political realignment, so that sensible people govern, and deal with real problems like the causes and effects of the climate crisis. But even in the best case a lot of people will suffer in the process. Those who will suffer as a result of default likely will include members of my family and myself. So I don't see any of this as a theoretical exercise.

  6. I know that people have short memories, but. First, the economic collapse. Despite Republican attempts to blame joblessness and slack economic growth on Obama, people still blame the Republicans. Next came ACA, and by the time the next election rolls around, I expect many people will feel they're better off than they were; at least able to keep kids on their insurance longer, able to get insurance if they've got pre-existing conditions, etc., and nary a death panel insight. While I doubt many voters will take the rolling CR's as a series of failures, all combined with the shutdown and talk of default?

    I don't know. That's asking people to forget a whole lot of stuff.

    I think there might be a third scenario in some districts: more RINO Republicans running as independent candidates, and either flipping the supposedly 'safe' districts to Democratic candidates in the general or outright winning as a saner alternative to people who come out of Tea Party Primaries.

    1. So then is this positing a "sore loser" effect, where more traditional Republican conservatives are challenged by Club for Growth conservatives, and then the loser of the primary continues on to the general to push a three-way scenario?

      That sounds like it would take a lot of money.

      Plus, you don't know if it wouldn't end up more like the governorship in Maine.

    2. Since I live in Maine, that's why it seems likely to happen; particularly if Republicans mostly stay home during the primaries and we get the worst of the worst of Tea Party challengers as the party nominees. I don't think the obvious independent-candidates likely to come out of the woodwork until the primary voters speak.

  7. You omit the question of what effect default would have on the electorate. Do you think Boehner and the GOP establishment are going to allow a vote at the last minute, making such an idea silly to consider? Or do you think that default, if it did happen, would have a different effect on the electorate?

    I'm leaning towards the first option, but I'd introduce another concept: if/when Boehner allows the vote and the debt ceiling is raised, I wonder if conservatives will then primary from the right the handful of GOPers who voted for the increase.

    That might take out incumbents (who tend to be more popular than the national party), and possibly introduce unreliable candidates. Provided that the Democratic Party succeeds in recruiting the strongest possible candidates in those districts, these factors might help tip a few more seats in the Democratic direction.

    In any case, these details can be easily incorporated into your analysis- basically, this shutdown/default threat isn't helping the GOP, but we shouldn't assume that it'll be fatal.


Note: Only a member of this blog may post a comment.

Who links to my website?